Airport parking space distribution method and system based on gridding time-space layering

Abstract

The invention provides an airport parking space allocation method and system based on gridding space-time layering, and the method comprises the steps: building an airport space-time grid model through obtaining and space-time correlation analysis of multi-source basic data, such as real-time flight dynamic data, parking space physical attributes, historical operation efficiency and ground guarantee resource states; then, on the basis of multi-level feature constraint fusion, adaptive weight dynamic adjustment and sparse constraint low-dimensional reconstruction, adaptive optimization facing complex space-time requirements and delay disturbance is achieved, and an airport parking space distribution matrix is obtained; generating an anti-conflict camera configuration set in combination with dynamic coupling and conflict verification of ground support resources; and finally, through physical constraint real-time inspection and historical efficiency mode matching, generating an optimal distribution matrix of the airports, so that the overall use efficiency and the operation guarantee capability of airport airport resources are remarkably improved, and the airport operation efficiency and the service quality are improved.

Description

Technical Field

[0001] The present invention relates to the field of airport operation scheduling, and in particular to an airport stand allocation method and system based on grid-based time-space stratification Background Art

[0002] Airport stand allocation is a crucial component of ensuring safe and efficient airport operations and management, and is essential foundational work in the air transportation process. A rational and accurate stand allocation plan can significantly improve the utilization efficiency of airport apron resources, reduce the risk of flight delays, ensure orderly flight operations, and enhance airport service quality and passenger satisfaction. Furthermore, a sound stand allocation strategy is directly linked to the efficient scheduling and coordination of airport ground support resources, avoiding service interruptions or delays caused by resource conflicts and congestion, and ensuring the overall efficient and smooth operation of the airport.

[0003] Currently, most airports rely on pre-defined rules or manual scheduling based on experience. This involves pre-planning slot allocations based on flight schedule information, simple manual rules, and static data, with temporary adjustments made based on human experience. Some airports have also begun to introduce computer-assisted decision-making, using simple mathematical models or single optimization algorithms to automatically pre-plan slot allocations based on basic flight schedules and static slot attributes. However, regardless of whether this approach relies on manual experience or traditional simple algorithms, most airports' current slot allocation methods still rely primarily on static pre-planning, with very limited real-time decision-making capabilities.

[0004] In summary, existing airport stand allocation schemes generally have obvious deficiencies and defects in actual operation. The main manifestations are that the stand allocation method is relatively rigid and lacks a rapid response mechanism to real-time flight dynamic changes and emergencies; airport operations involve a variety of support resources and complex constraints. Existing technologies have difficulty in conducting detailed and comprehensive analysis and decision-making in complex and changing environments, resulting in low airport resource utilization efficiency and long-term inefficient use of apron resources. In addition, existing schemes lack effective response measures when dealing with flight delays and resource conflicts, cannot achieve flexible and effective real-time optimization and scheduling, and fail to fully integrate successful experiences from historical operations, resulting in insufficient airport operation efficiency and reduced service quality, seriously affecting the overall airport operation support capabilities. Summary of the Invention

[0005] In view of the above actual situation, this application proposes an airport stand allocation method and system based on grid-based spatiotemporal stratification to solve the problems existing in the existing technology, such as high static degree of airport stand resource allocation, insufficient dynamic adaptability, lack of refined analysis and optimization means of multi-scale and multi-level spatiotemporal constraints, resulting in inefficient use of stand resources, insufficient delay response capability, frequent resource conflicts and insufficient integration of historical operating experience.

[0006] A method for allocating airport stands based on grid-based spatiotemporal stratification, comprising the following steps:

[0007] S1. Obtaining basic airport operation data, including real-time flight dynamics data, aircraft stand physical attribute data, historical operation performance data, historical flight delay records, ground support resource topology data, airport physical topology data, and ground support resource status data. The real-time flight dynamics data includes planned schedules, delay status, aircraft model parameters, and transfer connection requirements. The ground support resource topology data includes structured description data of the physical connection relationships and spatial layout between support facilities. The ground support resource status data includes dynamic monitoring data on the real-time occupancy status, movement trajectory, and service load of support equipment.

[0008] S2, performing spatiotemporal correlation analysis and grid modeling on the basic data to construct an airport spatiotemporal grid model, wherein the grid modeling includes spatiotemporal demand weight calculation and grid optimization fitting;

[0009] S3, performing adaptive optimization processing on the airport space-time grid model by coupling multi-level feature analysis with dynamic delay disturbances to obtain the airport stand distribution matrix. The adaptive optimization processing includes multi-level feature constraint fusion, dynamic adaptive weight adjustment, and low-dimensional reconstruction optimization based on sparse constraints.

[0010] S4, dynamically couples and optimizes the airport stand distribution matrix and ground support resource status data, and performs anti-conflict verification to generate an anti-conflict stand configuration set;

[0011] S5, perform real-time verification of physical constraints and optimization of historical performance patterns on the conflicting aircraft configuration set to generate the optimal aircraft distribution matrix.

[0012] Furthermore, the step S2 includes the following sub-steps:

[0013] S201, performing spatial correlation analysis on the physical attribute data of the aircraft stand and the topological data of the ground support resources, and generating an aircraft stand interaction network through adaptive scene mapping, wherein the spatial correlation analysis is a spatiotemporal adjacency calculation of the aircraft stand based on multi-scale topological aggregation;

[0014] S202: Calculate spatiotemporal demand weights based on the stand interaction network and real-time flight dynamic data, and generate a dynamic demand density matrix through asynchronous spatiotemporal grid mapping. The spatiotemporal demand weight calculation is based on weight allocation of flight dynamic priorities and time window conflict.

[0015] S203, performing spatiotemporal optimization fitting on the dynamic demand density matrix and historical operational efficiency data to generate an airport spatiotemporal grid model. The spatiotemporal optimization fitting is based on the aircraft stand occupancy rate and connection time distribution in the historical operational efficiency data, and the grid density is adjusted using a local sparse enhancement algorithm.

[0016] Furthermore, the S3 step includes the following sub-steps:

[0017] S301: Perform multi-level feature analysis and constraint fusion on the airport space-time grid model to generate a hierarchical optimization tensor. This hierarchical feature analysis and constraint fusion includes regional coordination matrix modeling at the international / domestic coordination layer, aircraft type-stand bipartite graph matching constraint generation at the wide-body / narrow-body adaptation layer, and temporal network flow analysis at the transfer connection optimization layer. A multi-dimensional optimization space is constructed through tensor superposition and hierarchical weight fusion.

[0018] S302: Adaptively adjust the weights of the hierarchical optimization tensor and the flight delay time series data to generate an optimization constraint space; the adaptive weight adjustment is based on time series delay disturbance analysis. The delay time series data is obtained by performing sliding window aggregation and propagation chain inference processing on the current delay status in the real-time flight dynamic data and historical flight delay records;

[0019] S303, performing adaptive tensor decomposition and low-dimensional solution space reconstruction processing on the optimized constraint space to obtain the airport stand distribution matrix, wherein the adaptive tensor decomposition and low-dimensional reconstruction processing include adaptive sparse constraint application based on dynamic scenes and low-rank representation of dynamic threshold screening.

[0020] Furthermore, the S4 step includes the following sub-steps:

[0021] S401, performing a dynamic coupling feasibility analysis on the airport stand distribution matrix and the real-time status of ground support resources, generating a stand-resource joint feasible domain matrix through a spatiotemporal constraint network. The dynamic coupling feasibility analysis is based on conflict detection and capacity matching between the physical topology constraints and the real-time resource load matrix;

[0022] S402, performing multi-objective dynamic anti-conflict optimization on the joint feasible domain matrix of the stand and resource to generate an anti-conflict stand configuration set, wherein the multi-objective dynamic anti-conflict optimization is based on a linear combination scoring function of the conflict resolution weight, the resource balancing factor and the flight priority vector, and is implemented through sparse solution space projection and local neighborhood search.

[0023] Furthermore, the step S5 includes the following sub-steps:

[0024] S501, performing real-time spatial conflict detection on the conflicting aircraft stand configuration set and the airport physical topology data, and generating a physically feasible configuration set through dynamic geometric boundary verification. The real-time spatial conflict detection is based on aircraft stand safety spacing constraints and vehicle path width threshold analysis.

[0025] S502, performing time-effectiveness pattern matching on the physical feasible configuration set and the historical operating efficiency data, and generating an optimal aircraft position distribution matrix by sliding window similarity weighting. The time-effectiveness pattern matching is based on the optimal allocation pattern extraction and priority fusion under historical scenarios of the same type.

[0026] Furthermore, the physical topology constraints in step S401 are obtained by constructing an adjacency matrix and performing graph theory connectivity analysis on the airport stand distribution matrix, and the real-time resource load matrix is ​​obtained by performing time series discretization and sliding window aggregation on the real-time status of ground support resources.

[0027] Furthermore, the conflict resolution weight in step S402 is obtained by dynamic statistical calculation of the conflict mark density in the joint feasible domain matrix of the aircraft slot and resource, the resource balancing factor is obtained by normalized analysis of the distribution entropy value of the real-time resource load matrix, and the flight priority vector is generated by static weight assignment of flight attributes.

[0028] Furthermore, the real-time spatial conflict detection in step S501 is based on the aircraft position safety distance constraint and the vehicle path passage width threshold analysis, including the aircraft position dynamic geometric boundary construction processing, real-time distance safety verification processing and vehicle path width verification processing.

[0029] Furthermore, the temporal pattern matching in step S502 is based on the optimal allocation pattern extraction and priority fusion under historical scenes of the same type, including historical scenes of the same type retrieval processing, allocation pattern similarity quantification processing, sliding window similarity weighting processing and comprehensive priority fusion processing.

[0030] In addition, the present application also discloses an airport stand allocation system based on grid-based spatiotemporal stratification, characterized in that the system includes:

[0031] an acquisition unit, configured to acquire basic airport operation data, including real-time flight dynamics data, aircraft stand physical attribute data, historical operation performance data, historical flight delay records, ground support resource topology data, airport physical topology data, and ground support resource status data. The real-time flight dynamics data includes planned schedules, delay status, aircraft model parameters, and transfer connection requirements. The ground support resource topology data includes structured description data of the physical connection relationships and spatial layout between support facilities. The ground support resource status data includes dynamic monitoring data on the real-time occupancy status, movement trajectory, and service load of support equipment.

[0032] A grid modeling unit is used to perform spatiotemporal correlation analysis and grid modeling on basic data to construct an airport spatiotemporal grid model. The grid modeling includes spatiotemporal demand weight calculation and grid optimization fitting.

[0033] An adaptive optimization unit is used to perform adaptive optimization processing on the airport space-time grid model by coupling multi-level feature analysis with dynamic delay disturbances to obtain the airport stand distribution matrix. The adaptive optimization processing includes multi-level feature constraint fusion, dynamic adjustment of adaptive weights, and low-dimensional reconstruction optimization based on sparse constraints.

[0034] The anti-conflict coupling unit is used to dynamically optimize the coupling of the airport stand distribution matrix and the ground support resource status data, and perform anti-conflict verification to generate an anti-conflict stand configuration set;

[0035] The configuration verification and result output unit is used to perform real-time verification of physical constraints and historical performance mode optimization on the conflicting aircraft configuration set to generate the optimal aircraft distribution matrix.

[0036] The present application proposes an airport stand allocation method and system based on grid-based spatiotemporal stratification, which realizes real-time dynamic configuration optimization of airport stand resources, multi-scale spatiotemporal constraint analysis, effective resolution of multi-resource conflicts, rapid adaptive adjustment under delay disturbance scenarios, and efficient integration of historical operating experience and real-time allocation plans, thereby significantly improving the overall utilization efficiency and operation guarantee capability of airport stand resources, and enhancing airport operation efficiency and service quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a flowchart of a method for allocating airport stands based on grid-based spatiotemporal stratification proposed in this application;

[0038] Figure 2 This is a flow chart of constructing an airport space-time grid model in an airport stand allocation method based on grid-based space-time stratification proposed in this application;

[0039] Figure 3This is a flow chart of generating an airport stand distribution matrix in an airport stand allocation method based on grid-based spatiotemporal stratification proposed in this application;

[0040] Figure 4 This is a flow chart of generating an anti-conflict aircraft stand configuration set in a method for airport aircraft stand allocation based on grid-based spatiotemporal stratification proposed in this application;

[0041] Figure 5 This is a flow chart of generating an optimal aircraft stand distribution matrix in an airport stand allocation method based on grid-based spatiotemporal stratification proposed in this application;

[0042] Figure 6 A schematic diagram of the structure of an airport stand allocation system based on grid-based spatiotemporal stratification provided in an embodiment of the present application; DETAILED DESCRIPTION

[0043] The following will clearly and completely describe the simulation technology route in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0044] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0045] The features and performance of the present invention are further described in detail below with reference to the embodiments. Figure 1 As shown, a method for allocating airport stands based on grid-based spatiotemporal stratification includes the following steps:

[0046] S1. Obtaining basic airport operation data, including real-time flight dynamics data, aircraft stand physical attribute data (including hierarchical labels of the aircraft stand region (international / domestic) and aircraft type (wide-body / narrow-body), supporting hierarchical constraint modeling), historical operation efficiency data, historical flight delay records, ground support resource topology data (including physical connection relationships between support facilities (gate-bridge mapping matrix, gate-baggage carousel connection diagram, shuttle bus connection path network, refueling truck path network topology diagram), spatial layout (special vehicle operation area boundary)), airport physical topology data, and ground support resource status data (including real-time occupancy status of support equipment (bridge occupancy rate, check-in counter utilization rate), movement trajectory (real-time shuttle bus location), and service load (ground service personnel workload)). The real-time flight dynamics data includes the planned schedule, delay status, aircraft type parameters, and transfer connection requirements. The ground support resource topology data includes structured description data of the physical connection relationship and spatial layout between support facilities. The ground support resource status data includes dynamic monitoring data of the real-time occupancy status, movement trajectory, and service load of support equipment.

[0047] In some embodiments, real-time flight dynamics data refers to a flow of flight operation information received and recorded in real time by the airport operation control center information platform during flight operations. In this embodiment, it specifically includes the flight schedule, current delay status, aircraft model parameters, and transfer connection requirements. The flight schedule consists of scheduled flight arrival and departure times and is updated in real time at a granularity of seconds or minutes based on the flight operation information system. The delay status specifically refers to the current flight delay duration and delay level, which is generated through sliding window aggregation and inference processing of the flight operation status data source and the flight delay database. The aircraft model parameters indicate whether the specific aircraft model belongs to the wide-body or narrow-body classification and its respective spatial dimension parameter information, which is obtained by real-time docking with the aircraft model parameter database provided by the airport maintenance department. Transfer connection requirements represent the transfer requirements and transfer time requirements for transfer passengers between a flight and subsequent flights, which are pushed and updated in real time by the passenger transfer connection data platform.

[0048] In some embodiments, the physical attribute data of the aircraft stand is obtained from a database that is pre-surveyed and regularly maintained and updated by the airport infrastructure management department. In this embodiment, it includes a hierarchical label of the region to which the aircraft stand belongs (international region or domestic region) and a label of the aircraft type suitable for the aircraft stand (suitable for wide-body aircraft or narrow-body aircraft). Among them, the region label to which the aircraft stand belongs is used to generate the regional coordination matrix for the collaborative layered optimization of international and domestic flights, and the aircraft type suitable for the aircraft stand label is used to construct the bipartite graph matching constraint of the aircraft type and the aircraft stand adaptation layer to support subsequent hierarchical constraint modeling.

[0049] Historical operational performance data is a data set formed by statistical analysis of the airport's previous flight operation status and aircraft stand occupancy status, including historical aircraft stand utilization rate, historical aircraft stand occupancy time, and historical transfer connection time distribution. The data set is obtained after the airport operation analysis system aggregates, mines and statistically analyzes historical flight data.

[0050] The historical flight delay records are composed of historical flight delay event records stored in the airport flight operation database. Each record clearly specifies the flight information where the delay event occurred, the duration of the delay, the cause of the delay, and the historical impact range. In this embodiment, historical delay statistics are generated by extracting from the delay database.

[0051] Ground support resource topology data describes the structured information of the physical connection relationship and spatial layout between various types of ground support facilities in the airport. Specifically, in this embodiment, the physical connection relationship between support facilities is represented by a structured data set, including the boarding gate-bridge mapping matrix, the boarding gate-baggage carousel link diagram, the shuttle bus connection path network and the refueling truck path network topology diagram. These data are stored in a matrix or graph structure to clearly define the connection relationship, path direction and spatial distance parameters of the support facilities. In addition, the spatial layout data describes the boundary range of the special vehicle operation area in the airport and is stored in the form of spatial coordinates and boundary contour polygons. The role of this type of topological data is to provide a clear constraint basis for the subsequent accessibility analysis and spatiotemporal coupling optimization between ground support resources and aircraft stands. The physical topology data of the airport are the physical coordinates of the aircraft stands, taxiway width, apron geometric boundaries, and the spatial geometric attributes of the safety interval buffer zone.

[0052] Ground support resource status data reflects the real-time operating status of airport ground support equipment and personnel. In some embodiments, it is collected and generated in real time through the airport operation monitoring system and Internet of Things sensors. In this embodiment, the status data includes a data set of three dimensions: the real-time occupancy status of the support equipment, the movement trajectory, and the service load. Among them, the real-time occupancy status of the support equipment is specifically manifested as real-time data such as the occupancy rate of the jet bridge and the utilization rate of the check-in counter, which are generated in real time by the equipment status monitoring system in the form of state occupancy ratio or occupancy time; the movement trajectory data is specifically manifested as the real-time position of the shuttle bus, which is generated by real-time positioning of the global positioning system device installed on the shuttle bus; the service load data is specifically manifested as the workload level of the ground service personnel, which is generated in real time through the airport personnel scheduling and management system to clarify the personnel load and service saturation. These status data are used to analyze the use of ground support resources in real time, so as to achieve dynamic coupling optimization of the aircraft positions and support resource status in subsequent steps.

[0053] S2, performing spatiotemporal correlation analysis and grid modeling on the basic data to construct an airport spatiotemporal grid model, wherein the grid modeling includes spatiotemporal demand weight calculation and grid optimization fitting processing;

[0054] For details, please refer to the attached Figure 2 As shown, this step includes the following sub-steps:

[0055] S201, performing spatial correlation analysis on the physical attribute data of the aircraft stand and the topological data of the ground support resources, and generating an aircraft stand interaction network through adaptive scene mapping, wherein the spatial correlation analysis is a spatiotemporal adjacency calculation of the aircraft stand based on multi-scale topological aggregation;

[0056] In some embodiments, spatial association analysis is based on the calculation of the spatiotemporal adjacency of aircraft stands based on multi-scale topological aggregation. In this embodiment, multi-scale topological aggregation means that the spatial proximity relationship between aircraft stands and ground support facilities in the airport is expressed in a multi-scale hierarchical structure, that is, it includes fine-grained local topological relationships and coarse-grained overall topological structures to fully reflect the spatial layout characteristics of airport resources. Specifically, this embodiment obtains topological adjacency matrices of different granularity scales by performing a mapping relationship analysis on the physical attribute data of the aircraft stands and the topological data of the ground support resources. These matrices mathematically define the spatial adjacency or path accessibility relationship between aircraft stands and between aircraft stands and ground support facilities.

[0057] In this embodiment, the physical attribute data of the aircraft stand includes the aircraft stand location coordinates, the aircraft type label (wide-body or narrow-body) and the region label (international or domestic); the ground support resource topology data includes the gate-bridge mapping matrix, the gate-baggage carousel link diagram, the shuttle bus connection path network and the refueling truck path network topology diagram, as well as the boundary coordinates of the special vehicle operation area. Based on this data, this embodiment uses a spatial topology analysis algorithm to define the spatial connections between each facility node in a graph theory manner, and constructs a complete airport facility-aircraft stand topology network diagram G = (V, E), where the node set V = {V s ,V p} By facility node V s (including boarding gate, jet bridge, baggage carousel, special vehicle area nodes) and aircraft stand node V p Composition: The edge set E describes the path connection relationship between each node, and each edge expresses the spatial distance and connection state characteristics between two nodes in a weighted manner.

[0058] In this embodiment, the mathematical model defining the spatial relationship between nodes is expressed as an adjacency matrix A. Assuming that the total number of airport facility and aircraft stand nodes is n, the matrix is ​​an n×n square matrix, and its elements are defined as: Where d(i, j) represents the actual measured physical length or spatial distance of the path between node i and node j. If there is no direct path, it is set to infinity.

[0059] In some implementations, this embodiment utilizes the above-mentioned adjacency matrix A to further perform multi-scale topological aggregation calculations, thereby obtaining a spatiotemporal adjacency matrix that reflects the spatial relationship between adjacent aircraft positions. The specific calculation method is as follows: Set the spatial scale aggregation function f σ (·), defines the scale-space aggregation adjacency strength between camera positions for: Where p and q represent any two camera nodes, and the parameter σ is a scale space parameter that controls the spatial granularity of the topological relationship. Different values ​​of σ constitute a multi-scale topological expression.

[0060] In this embodiment, after calculating the spatial adjacency strength at different scale parameters using the scale space aggregation function, the topology of the aircraft stand interaction network is generated based on the adaptive scene mapping technology. That is, the scale parameter combination suitable for the current scenario is automatically selected based on the actual operation scenario and spatial layout of the airport. The mathematical expression of adaptive scale selection is: Where C pq This represents the reference value of the stand spatial correlation characteristics obtained from actual scenarios, generated by statistical aggregation of historical data. This adaptive scale selection method ensures that the stand interaction network can optimally reflect the actual airport operation space characteristics.

[0061] In this embodiment, the generated machine position interaction network is denoted as G pos =(V p ,E pos ), where E pos The weight w pq It is defined as the optimal value of the position adjacency strength under the above adaptive aggregation scale: In this embodiment, the aircraft stand interaction network constructed through this process will clearly represent the spatial topological relationship, thereby providing effective data support for the subsequent calculation of spatiotemporal demand weights and the construction of the airport spatiotemporal grid model.

[0062] S202: Calculate spatiotemporal demand weights based on the stand interaction network and real-time flight dynamic data, and generate a dynamic demand density matrix through asynchronous spatiotemporal grid mapping. The spatiotemporal demand weight calculation is based on weight allocation of flight dynamic priorities and time window conflict.

[0063] In some embodiments, step S202 includes four processes: flight dynamic priority determination, time window conflict quantification, stand adjacency conflict propagation, and spatiotemporal mapping fusion. Flight dynamic priority determination is a quantitative assignment based on the delay level and transfer connection urgency in real-time flight dynamic data. The delay level in this embodiment is a numerical value obtained by quantifying the delay status in real-time flight dynamic data, which is used to represent the delay severity of the flight at the current operating time. Time window conflict quantification counts overlapping conflicts between flights in the form of discrete time windows. Stand adjacency conflict propagation utilizes the adjacency weights of the stand interaction network to diffuse or attenuate flight conflicts within a spatially adjacent range. Spatiotemporal mapping fusion constructs a demand density matrix within different time windows based on the weighted results.

[0064] In this embodiment, the flight set is defined as F = {f1, f2, ..., f m}, the camera position set is defined as G={g1,g2,…,g l}, the time window set is defined as T = {t1, t2, ..., t n For flight f i , the delay level quantization value is recorded as δ i , the transit connection urgency quantification value is recorded as γ i , both are extracted from real-time flight dynamic data, and the flight priority vector P = {p1, p2, ..., p m}Element p in i :p i =α·δ i +β·γ i , symbols α and β are weight coefficients, which are configured by the airport operation management strategy, δ i with γ i The larger the value, the higher the priority of the flight's demand for slot resources.

[0065] In this embodiment, the time window conflict quantification is based on the statistics of the overlap of flights within the time window. k , define the time window conflict matrix in This matrix describes the competition between any two flights in the same time window.

[0066] In this embodiment, in order to express flight conflicts more accurately in the stand space, the stand interaction network G is used. pos =(V p ,E pos ), where V p is the set of camera nodes, E pos Contains the adjacency relationship between camera pairs, camera pairs (g p ,gq ) is recorded as w p,q Through the propagation of the aircraft position adjacent conflict, the time window conflict degree is coupled with the aircraft position spatial adjacency to obtain the aircraft position adjacent conflict factor The specific process is to Perform weighted diffusion or attenuation in the stand network space to strengthen the conflicts between flights participating in the same stand or adjacent stands. Let Λ be the stand adjacency propagation operator, then This operator propagates the flight conflict intensity between adjacent stands. If two stands g p With g q The adjacency weight w p,q If it is larger, then flight f i With f j The intensity of competition between these two slots is accordingly considered to be more significant.

[0067] In this embodiment, the spatiotemporal mapping fusion is to comprehensively calculate the flight priority and the slot adjacency conflict factor to obtain the weighted result of the flight's demand for slots in each time window. i In the time window t k The basic demand weight calculation is defined as: Where η is the conflict amplification factor, The conflict intensity between flight fi and flight fj after propagation through the stand interaction network is shown in FIG. In this embodiment, a discrete space-time grid set Ω is defined as: s ,τ e ,x}, where τ s ,τ e ∈T represents the start and end time of the time interval, and x∈G represents the aircraft stand space node. The dynamic demand density matrix D is obtained by summing up the flight demand weights in different time periods and different aircraft stand nodes. s ,τ e , the demand density on x) is recorded as The calculation method is: φ(f i ,x,τ s ,τ e ), is a binary function, representing flight f i In the time interval [τ s ,τ e ] whether there is demand for camera position x; if occupied, it is recorded as 1, otherwise it is recorded as 0. In some embodiments, based on the adjacency weights of the camera interaction network, the dynamic demand density matrix D can also be coupled with the adjacency propagation operator to achieve smoothing or attenuation of the impact of adjacent camera positions, providing more detailed spatial correlation information for subsequent spatiotemporal grid model construction.

[0068] S203, performing spatiotemporal optimization fitting on the dynamic demand density matrix and historical operational efficiency data to generate an airport spatiotemporal grid model. The spatiotemporal optimization fitting is based on the aircraft stand occupancy rate and connection time distribution in the historical operational efficiency data, and the grid density is adjusted using a local sparse enhancement algorithm.

[0069] In some embodiments, the S203 step process includes joint modeling of historical occupancy and connection time, construction of spatiotemporal difference measure and objective function, execution of local sparse enhancement algorithm, and output of airport spatiotemporal grid model. In this embodiment, the historical operational performance data includes historical aircraft stand occupancy sequence and historical connection time distribution. The historical aircraft stand occupancy sequence records the occupancy of past flights at each aircraft stand in a discrete time window manner to form an occupancy function R(g,t), where g∈G represents the aircraft stand and t∈T represents the discrete time window; the historical connection time distribution characterizes the probability distribution of passenger transfer connection time, denoted as H(τ), where τ represents the transfer connection time. The above historical performance data is combined with the dynamic demand density matrix to construct a spatiotemporal difference measure to quantify the degree of deviation between the current demand density and the historical performance pattern. The spatiotemporal difference function is defined as: where d t,g is the element of the demand density matrix D at time window t and aircraft position g, R(g,t) is the statistical value of the historical aircraft position occupancy rate sequence at the corresponding time window and aircraft position, is the deviation measurement coefficient.

[0070] In this embodiment, in order to reflect the distribution characteristics of transfer connection time in the airport space-time grid model, a constraint based on the connection time distribution is introduced. This constraint is quantified by analyzing the coupling degree between the temporal distribution of demand density and the historical connection time distribution. in is the connection time distribution constraint coefficient, Γ(D,τ) is the flight connection time distribution function based on the demand density matrix, and is obtained by counting the time intervals between adjacent flights at corresponding stands.

[0071] In some embodiments, the local sparsity enhancement algorithm adaptively adjusts the local area of ​​the demand density matrix where there is a large deviation from the historical performance data. The algorithm achieves grid density correction by detecting sparsity differences within the local neighborhood and applying a sparsity penalty term. Let the local neighborhood N(g) represent the set of adjacent stations of station g in the station interaction network. The local sparsity penalty term is introduced and is defined as Where ζ is the sparsity adjustment index, is the sparse enhancement coefficient, d t,g with d t,hThe demand density of position g and its neighboring position h within time window t is penalized for extremely unbalanced or overly concentrated demand distribution in a local area, guiding the model to more closely follow the historical operating efficiency distribution pattern.

[0072] In this embodiment, the above-mentioned spatiotemporal difference measure, connection time distribution constraint and local sparse penalty term are integrated to construct an overall objective function: By dynamically adjusting the element values ​​of D during the iteration process, the minimization solution is cyclically executed and convergence criteria are met, and the revised demand density matrix is ​​output. After the iteration is completed, the discrete space-time grid where the demand density matrix is ​​located is combined with the aircraft stand space labels and time windows to generate an airport space-time grid model that includes density distribution, historical efficiency adaptability, and connection time constraint information.

[0073] S3, performing adaptive optimization processing on the airport space-time grid model by coupling multi-level feature analysis with dynamic delay disturbances to obtain the airport stand distribution matrix. The adaptive optimization processing includes multi-level feature constraint fusion, dynamic adaptive weight adjustment, and low-dimensional reconstruction optimization based on sparse constraints.

[0074] For details, please refer to the attached Figure 3 As shown, this step includes the following sub-steps:

[0075] S301: Perform multi-level feature analysis and constraint fusion on the airport space-time grid model to generate a hierarchical optimization tensor. This hierarchical feature analysis and constraint fusion includes regional coordination matrix modeling at the international / domestic coordination layer, aircraft type-stand bipartite graph matching constraint generation at the wide-body / narrow-body adaptation layer, and temporal network flow analysis at the transfer connection optimization layer. A multi-dimensional optimization space is constructed through tensor superposition and hierarchical weight fusion.

[0076] In this embodiment, the regional coordination matrix modeling of the international / domestic coordination layer is based on the aircraft stand area label information in the airport space-time grid model. Define the aircraft stand set G = {g1, g2, ..., g l}, each camera position g i Divide into two categories of regions: international or domestic, and construct the regional coordination matrix S based on this region =[s i,c ]. The symbol c∈{I nt,Dom} represents the region classification, s i,c Indicates camera position g i The fitness or feasibility quantitative value under the corresponding regional classification. Introducing regional collaborative weight ω region In order to balance the distribution preferences at the international and domestic levels, the constraints on regional coordinated distribution will be reflected through weight fusion when multiple levels are superimposed.

[0077] In some embodiments, the aircraft model-slot bipartite graph matching constraint of the wide / narrow adaptation layer is composed of a set of aircraft models and a set of slots forming a bipartite graph B = (U, V, E), where U represents an aircraft model with a wide or narrow body identifier, V represents a slot that can accommodate the corresponding aircraft model, and E represents the adaptation relationship between the aircraft model and the slot. Define the adjacency matrix M of the bipartite graph type =[m u,v ],in The feasibility of matching between aircraft type and position is reflected through the adjacency matrix. Introducing matching variable x u,v ∈{0,1}, and define the matching constraints In the subsequent multi-dimensional optimization space, M type The constraint weight ω of the wide-body / narrow-body adaptation layer is matched with the constraint type Fusion with other level features.

[0078] In this embodiment, the time network flow analysis of the transfer connection optimization layer is based on the order of flights occupying the aircraft seats in the time-space grid and the connection time requirements of flight transfer passengers, and the path and flow constraints of the connection process are described in the form of network flow. A directed graph D = (N, A) is constructed, where N represents the set of nodes jointly identified by the flight occupation period and the aircraft seat location, A represents the set of possible connection paths, and the flow rate f a represents the transfer passenger flow on arc a. The flow balance constraint is executed on all nodes: And set the capacity u on each arc a a The capacity constraint of the network flow is formed by limiting the maximum number of transfer passengers that can pass through: The connection time threshold is quantified by the time interval corresponding to arc a. If the time interval is shorter than the shortest time required for transit connection, the capacity value of arc a is set to 0, thereby eliminating possible paths that do not meet the connection requirements at the network level. Based on the above time network flow analysis, the connection constraint matrix M is constructed. flow , and introduce the connection layer weight ω flow The optimization objective is incorporated into the multidimensional space.

[0079] In this embodiment, S is superimposed by tensor region 、M type With M flow Placed in a unified multi-dimensional structure, the hierarchical optimization tensor Τ=Τ(S region ,M type ,M flow ), and combined with the level weight ω through the following objective function region ,ω type ,ω flow Realize multi-dimensional optimization space construction: Functions f(·), g(·), and h(·) perform structured metrics or matching metrics on the three types of feature matrices, respectively. In subsequent steps, this hierarchical optimization tensor will be combined with real-time delay perturbation adjustments and other constraints to further perform a multidimensional solution space search, thereby finding an appropriate airport stand allocation solution under multiple objectives and multi-level requirements.

[0080] S302: Adaptively adjust the weights of the hierarchical optimization tensor and the flight delay time series data to generate an optimization constraint space; the adaptive weight adjustment is based on time series delay disturbance analysis. The delay time series data is obtained by performing sliding window aggregation and propagation chain inference processing on the current delay status in the real-time flight dynamic data and historical flight delay records;

[0081] In some embodiments, the process incorporates the evolution characteristics of flight delays in the time dimension into a multi-level constraint system based on timing delay disturbance analysis, and forms an optimization constraint space that can dynamically adapt to delay changes by adjusting the weight parameters of each level in the hierarchical optimization tensor.

[0082] In this embodiment, the layered optimization tensor is recorded as T = T (S region ,M type ,M flow ), where S region 、M type With M flow They respectively represent the regional coordination matrix modeling results of the international / domestic coordination layer, the aircraft type-stand bipartite graph matching relationship of the wide-body / narrow-body adaptation layer, and the temporal network flow analysis results of the transfer connection optimization layer. The hierarchical optimization tensor is fused using multiple weight parameters. Flight delay time series data is obtained by combining the current delay status in real-time flight dynamics data with records in the historical delay database through sliding window aggregation and propagation chain inference. To promptly reflect the impact of delays in the hierarchical optimization tensor, in this embodiment, a time series delay disturbance factor Δ(t) is defined to measure the cumulative degree and propagation risk of flight delays at time t.

[0083] In this embodiment, the calculation process of the timing delay disturbance factor Δ(t) includes two parts: sliding window aggregation and propagation chain reasoning. For the recent A sliding window aggregation function is defined for a series of delay measurement data collected within a time slice (such as the number of delayed flights, cumulative delay duration, or weighted sum of delay levels). This function outputs the current time t before and after The average delay measure within a time slice. Propagation chain reasoning is based on the historical pattern of delay propagation and the current airport operation status. By identifying the probability of delay chains occurring in the sliding window sequence, the delay propagation function is defined: where σ krepresents the impact of the delay in the kth time slot on subsequent time slots, and ρ(·) is the delay propagation assessment function. A weighted combination of W(t) and Γ(t) yields the timing delay disturbance factor: Δ(t) = α1W(t) + α2Γ(t), where α1 and α2 are adjustment coefficients used to balance the contribution of average delay and propagation chain inference. A larger value of this disturbance factor indicates a higher degree of delay and higher risk of propagation.

[0084] In this embodiment, in order to incorporate the timing delay disturbance factor Δ(t) into the weight adjustment of the hierarchical optimization tensor, an adaptive weight adjustment function Φ is defined. w Right region ,ω type ,ω flow Dynamic update. The specific weight value after adaptive adjustment at time t is as follows: Symbol β region ,β type ,β flow is the delay sensitivity coefficient of the corresponding level, which is used to determine the impact of delay disturbance on the weight of each level. The new weight set is obtained by updating the above weights at time t: In this embodiment, the updated Ω * (t) is combined with the hierarchical optimization tensor T to form the optimization constraint space Z(t), which is denoted as in By integrating the tensor representation of the hierarchical structure and weight parameters, the optimization constraint space represented by Z(t) can be dynamically adjusted as the flight delay disturbance factor changes. Through this adaptive weight adjustment mechanism, the airport stand allocation algorithm in this embodiment retains the underlying multi-level constraint structure while also providing flexible response to real-time delay conditions.

[0085] S303, performing adaptive tensor decomposition and low-dimensional solution space reconstruction on the optimized constraint space to obtain the airport stand distribution matrix, wherein the adaptive tensor decomposition and low-dimensional reconstruction include applying adaptive sparse constraints based on dynamic scenarios and low-rank representation of dynamic threshold screening.

[0086] In some embodiments, by applying dynamic scene-driven sparse constraints in the optimization constraint space Z(t) and combining it with a low-rank representation of dynamic threshold screening, the high-dimensional structure of the tensor is compressed and decomposed to obtain a solution space that satisfies multiple constraints in both time and space dimensions.

[0087] In this embodiment, let Represents the three-dimensional optimization constraint tensor generated in the previous step. Dimensions I and J correspond to the slot hierarchy and flight attribute mapping, and dimension K corresponds to time or connection layer expansion. To achieve low-rank decomposition, the following tensor decomposition model is defined: in represents the outer product operation, R is the set decomposition rank, λ r represents the corresponding scalar coefficient in the decomposition, are the factor vectors corresponding to the rth component in three dimensions. This decomposition represents Z(t) with a finite number of components, thereby summarizing spatiotemporal relationships and multi-level constraint information in the latent dimensions.

[0088] In some embodiments, the adaptive sparsity constraint is implemented by decomposing the coefficients λ r or factor vector This is achieved by applying a sparse regularization term on it. Let ||·|| represent the vector l1 norm and introduce a sparse penalty term Where β is a sparsity penalty coefficient, which constrains the components of the factor vector to eliminate features that have no significant impact at the local level in dynamic scenarios and attenuate or suppress redundant dimensions of the hierarchical optimization tensor. This sparsity constraint adaptively adjusts the value of β during periods of delay disturbances or drastic load changes, providing flexible control capabilities for tensor decomposition.

[0089] In this embodiment, the low-rank representation of dynamic threshold screening is based on or factor vector The gradual screening and reduction process. Let θ(t) be the threshold function coupled with the timing delay disturbance factor Δ(t), when |λ r | or When it is lower than θ(t), the corresponding component is judged to have insufficient contribution to the current scene in the decomposition and is set to zero or eliminated. This process can be performed periodically in the decomposition algorithm iteration. Define the threshold judgment expression The dynamic screening results of λr are fed back to the main loop of the decomposition algorithm through this judgment function, forming a dual fusion of low-rank representation and sparse constraints.

[0090] In this embodiment, after the adaptive tensor decomposition and low-dimensional solution space reconstruction are completed, a solution that meets the multi-level constraint requirements on the time axis and the camera space axis is obtained, that is, the reconstruction form of the tensor X is Where R′≤R, The decomposed components are filtered by dynamic thresholds and corrected by sparsity constraints. Projecting this low-dimensional reconstruction onto the stand index and time dimensions yields the airport stand distribution matrix. This approach achieves dimensionality reduction for multi-dimensional optimization problems, ensuring the feasibility and flexibility of stand allocation solutions under multi-level coordination and delay disturbances.

[0091] S4, dynamically couples and optimizes the airport stand distribution matrix and ground support resource status data, and performs anti-conflict verification to generate an anti-conflict stand configuration set;

[0092] For details, please refer to the attached Figure 4 As shown, this step includes the following sub-steps:

[0093] S401: Perform a dynamic coupling feasibility analysis on the airport stand distribution matrix and ground support resource status data, generating a stand-resource joint feasible domain matrix through a spatiotemporal constraint network. The dynamic coupling feasibility analysis is based on conflict detection and capacity matching between physical topology constraints (stand-bridge mapping, shuttle bus path accessibility) and a real-time resource load matrix (bridge occupancy rate, ground support personnel load).

[0094] In some embodiments, the physical topology constraints are obtained by constructing an adjacency matrix and performing graph theory connectivity analysis on the airport stand distribution matrix; the real-time resource load matrix is ​​obtained by performing time series discretization and sliding window aggregation on ground support resource status data.

[0095] In some implementations, physical topology constraints are integrated with the real-time resource load matrix, and conflict detection and capacity matching are used to identify the feasibility of coupling between aircraft stands and various ground support resources. In this embodiment, the physical topology constraints are based on the spatial connection relationship of aircraft stands obtained by constructing the adjacency matrix and performing graph theory connectivity analysis on the airport aircraft stand distribution matrix. Let the aircraft stand set be denoted as G = {g1, g2, ..., g m The stand distribution matrix can be viewed as a mapping of each stand to the corresponding flight within a discrete time window. Based on the stand layout and resource topology, the adjacency matrix A = [a ij ],in In some embodiments, the aircraft stand-bridge mapping relationship is given by the actual physical docking information between the boarding gate and the aircraft bridge, and the shuttle bus path accessibility is given by the accessible route constraints in the shuttle bus connection network. The adjacency matrix is ​​expanded through connectivity analysis to obtain the connection mapping of aircraft stands and resource nodes in the spatiotemporal constraint network.

[0096] In this embodiment, the real-time resource load matrix is ​​obtained by performing time series discretization and sliding window aggregation on the time axis based on monitoring data such as the bridge occupancy rate, ground service personnel load, and shuttle equipment usage. n}, in the time window T k Define the resource load matrix in in Represents resource node r j In the time window T k The load or occupancy value of the resource node r. j Defining capacity thresholds If satisfied It is considered to be in the time window T kThe resource is available.

[0097] In some embodiments, conflict detection is performed by determining whether there are multiple aircraft positions competing for the same resource node within the same time window, or whether the resource load of a single aircraft position exceeds the capacity threshold at the same time. Let Φ(g i ,r j ,T k )∈{0,1}, indicating the camera position g i In the time window T k Is resource node required? j Support, if And a ij =1, resource conflicts may occur. At the same time, it is necessary to verify in Indicates camera position g i The amount of additional load introduced. If the capacity is exceeded, the conflict is considered unsolvable. If the inequality is satisfied, the capacity is acceptable.

[0098] In this embodiment, the above-mentioned conflict detection and capacity matching results are combined to obtain the joint feasible region matrix of the slot-resource The matrix elements When the machine position i and resources j In the time window T k If there is path connectivity and no overload conflict occurs, it is recorded as 1, otherwise it is recorded as 0. k The above analysis and calculation are performed to obtain the slot-resource joint feasible region matrix sequence covering the entire scheduling cycle, which is used in the subsequent anti-conflict optimization step.

[0099] S402, performing multi-objective dynamic anti-conflict optimization on the joint feasible domain matrix of the stand and resource to generate an anti-conflict stand configuration set, wherein the multi-objective dynamic anti-conflict optimization is based on a linear combination scoring function of the conflict resolution weight, the resource balancing factor and the flight priority vector, and is implemented through sparse solution space projection and local neighborhood search.

[0100] In some embodiments, the conflict resolution weight is obtained by dynamic statistical calculation of the conflict mark density in the slot-resource joint feasible domain matrix, the resource balancing factor is obtained by normalized analysis of the distribution entropy value of the real-time resource load matrix, and the flight priority vector is generated by static weight assignment of flight attributes (VIP identification, transfer urgency).

[0101] In this embodiment, the joint feasible matrix of the slot and resource is denoted as M feas ∈{0,1} m×n×T , where m is the number of slots, n is the number of resource nodes, and T is the number of discrete time windows. The elements of this matrix Mfeas (i, j, t) = 1 means that the camera position g is within the time window t. i and resources j The coupling between them is feasible, otherwise it is 0. To achieve multi-target anti-collision, three metrics are introduced in this embodiment: conflict resolution weight W c , resource balancing factor F b With the flight priority vector P f .

[0102] In some embodiments, the conflict resolution weight W c It is the numerical quantification result obtained by dynamic statistics of the conflict mark density in the feasible domain matrix. Let C count (i,j,t) represents the camera position g i In time window t, for resource r j The number of conflicts or conflict intensity is calculated. If the conflict is caused by multi-camera competition or by exceeding resource capacity, conflict markers are accumulated. The conflict marker density is defined as: By dividing ρ c (t) Perform smoothing or sliding window aggregation on the time axis to obtain the conflict resolution weight W c (t) = f c (ρ c (t)), where f c is the conflict weight mapping function, which makes the conflict resolution weight increase with the increase of conflict density.

[0103] In some embodiments, the resource balancing factor F b It is obtained by normalizing the distribution entropy value of the real-time resource load matrix. Represents a resource set {r1,…,r n In the load matrix of time window t, define the resource distribution entropy: in Represents resource node r j The proportion of the total load in the time window t. By mapping H in the normalized interval [0,1] b (t), and obtain the resource balancing factor: in and It is the minimum and maximum distribution entropy within the historical or forecast range, which is used to ensure that the equilibrium factor value is within a stable range.

[0104] In some embodiments, the flight priority vector P f Generated by mapping flight attribute information, including VIP identification, transfer urgency, etc. For the flight set {f1,…,f q}, define p i Indicates flight f iIf the flight has VIP identification or transit urgency, then let p i Improve accordingly. Let P f ={p1,…,p q} is a vector set consisting of all flight priorities.

[0105] In this embodiment, the score of each allocation scheme in the joint feasible matrix of slots and resources is recorded as Φ(i, j, t), and the following linear combination function Φ(i, j, k) = αW is used. c (t)+βF b (t)+γp f (i, t), where α, β, γ are weight coefficients, p f (i,t) represents the time window t at the aircraft position g i The priority value of flight f(i, t) is used to associate the priority relationship between slots and flights. A lower combined score indicates a higher conflict risk or a higher resource load imbalance, while a higher combined score indicates a more ideal allocation.

[0106] The sparse solution space projection is to prune a large number of zero-value or invalid allocation states in the feasible domain, and retain the allocation units with potential improvement space in the conflict resolution dimension or resource balance dimension. feas When projecting to several subspaces, for the allocation state (i, j, t), if Φ(i, j, t) is lower than the threshold, the allocation state is eliminated or placed in the priority optimization queue.

[0107] Local neighborhood search is based on fine-tuning the interactions within adjacent slots or time windows to gradually improve the overall score. The neighborhood N(i, j, t) is defined to include combinations of states that are adjacent or interchangeable with (i, j, t), such as spatially adjacent slots or temporally adjacent flight slots. Local search and swapping between Φ(i′, j′, t′) and Φ(i, j, t) are performed. If the swap improves the overall score, the allocation is changed, and the algorithm iterates to search for the global optimal or near-optimal solution.

[0108] In this embodiment, based on the above-mentioned multi-objective dynamic anti-conflict optimization process, the anti-conflict aircraft stand configuration set formed can seek a balance among the three requirements of resource balancing, conflict resolution and flight priority, providing a basic allocation plan for subsequent physical feasibility verification and historical performance model optimization.

[0109] S5, performs real-time verification of physical constraints and optimization of historical performance patterns against the conflicting aircraft configuration set to generate the optimal aircraft distribution matrix;

[0110] For details, please refer to the attached Figure 5 As shown, this step includes the following sub-steps:

[0111] S501, performing real-time spatial conflict detection on the conflicting aircraft stand configuration set and the airport physical topology data, and generating a physically feasible configuration set through dynamic geometric boundary verification. The real-time spatial conflict detection is based on aircraft stand safety spacing constraints and vehicle path width threshold analysis.

[0112] In some embodiments, the real-time spatial conflict detection is based on the aircraft stand safety spacing constraint and the vehicle path width threshold analysis, including aircraft stand position dynamic geometric boundary construction processing, real-time spacing safety verification processing, and vehicle path width verification processing. In this embodiment, the aircraft stand position dynamic geometric boundary construction processing uses the airport physical topology data to determine the aircraft stand position coordinates and establish the geometric boundary of the aircraft stand occupied space. For the aircraft stand set G = {g1, g2, ..., g m}, each camera g i The center position in the airport plane coordinate system is expressed as (x i ,y i ). Define the geometric boundary of the space occupied by the camera as a circular or polygonal area, and the radius of the circular area is recorded as r i , the polygonal area is composed of the coordinate point set {(x i,k ,y i,k )|k=1,2,…,K}. The mathematical expression of the space occupied by the camera is: ξ(g i )={(x,y)|(xx i ) 2 +(yy i ) 2 ≤r i 2} or P(g i )={(x,y)|(x,y) is located in the coordinate set {(x i ,y i ,k)}}, the boundary of the above area represents the actual occupied space of the camera position on the ground, which is used to determine the safety distance constraint between adjacent camera positions when implementing dynamic space conflict detection.

[0113] In this embodiment, the real-time distance safety check process performs real-time spatial distance detection on multiple aircraft positions allocated at the same time in the anti-conflict aircraft position configuration set to determine whether the aircraft position spatial position distribution meets the airport safety operation standards. i and g j , let their corresponding position coordinates be (x i ,y i ) and (x j ,y j ). Define the distance between the centers of the two cameras as the Euclidean distance d ijAirport safety standards require that the distance between the centers of two aircraft stands should not be less than the specified safety threshold d min , so the spacing safety constraint is expressed as: d ij ≥d min , if any camera pair (g i ,g j ) does not meet the above conditions, it is considered a spatial location conflict and is not included in the physically feasible configuration set.

[0114] In some embodiments, the vehicle path width verification process utilizes the shuttle bus, refueling truck, or baggage transfer vehicle path network topology data in the airport physical topology data to perform real-time detection of the vehicle path width in the aircraft stand configuration plan. The vehicle path network is represented by a graph structure R = (V, E, W), where the node set V is the vehicle path node, the edge set E is the passable path between nodes, and W(e) is the actual pass width corresponding to the path edge e∈E. Let the vehicle pass width threshold be W min , then the vehicle path width constraint is expressed as: In real-time spatial conflict detection, the space occupied by the aircraft is projected into the vehicle path network topology. If there is a path edge e′ that is invaded by the aircraft space and causes W(e′) < W min This means that this camera configuration plan will cause vehicle traffic conflicts at this path location and needs to be eliminated.

[0115] In this embodiment, based on the above-mentioned space safety distance constraint verification and vehicle path width threshold verification process, a physical feasibility identification function F is constructed. phy (g i ,t), used to identify the boarding position g at time t i Physical space feasibility of the configuration: The physical feasibility identification function values ​​of all time and all machine combinations are combined to form the physical feasible configuration set matrix M phy , whose matrix elements are defined as: m phy (i,t)=F phy (g i ,t) This matrix represents the feasibility status of the physical space of all schemes in the anti-conflict camera configuration set after dynamic geometric boundary verification, providing the necessary data basis and spatial constraint guarantee for subsequent historical operation efficiency pattern matching and the generation of the final optimal camera distribution matrix.

[0116] S502, perform time-sensitive pattern matching on the physical feasible configuration set and historical operational performance data, and generate an optimal stand distribution matrix through sliding window similarity weighting. The time-sensitive pattern matching is based on the optimal allocation pattern extraction and priority fusion under historical similar scenarios (flight combination, time period, delay level).

[0117] In this embodiment, the temporal pattern matching is based on the optimal allocation pattern extraction and priority fusion under historical similar scenes, including historical similar scene retrieval processing, allocation pattern similarity quantification processing, sliding window similarity weighting processing and comprehensive priority fusion processing.

[0118] In some embodiments, the historical scene retrieval process is based on the flight combination structure, flight operation period characteristics and flight delay level in the airport historical operation efficiency database, and identifies historical sample scenes with similar characteristics to the current scene through feature space distance measurement. The feature vector of the current scene to be matched is defined as v cur =(v1,v2,…,v p ), the historical sample scene set is H = {h1,h2,…,h N}, where each historical scene sample h k The corresponding eigenvector is v hk =(v k,1 ,v k,2 ,…,v k,p ). The feature space distance is defined as the Euclidean distance: By setting the distance threshold D thr , select the one that satisfies the condition D(v cur ,v hk )≤D thr The historical scenes are candidate sets of the same type of scenes

[0119] In this embodiment, the allocation pattern similarity quantification process performs a refined measurement of the similarity between the camera position distribution of the current physically feasible configuration set to be matched and the candidate historical scene samples. Suppose the camera position distribution matrix of the current to-be-matched solution is M cur ∈{0,1} m×T , historical sample scene h k The corresponding camera distribution matrix is ​​M hk ∈{0,1} m×T Then define the distribution pattern similarity function: The above formula is a cosine similarity calculation method, and the value range is between [0,1]. The higher the value, the more similar the current plan is to the historical sample allocation pattern.

[0120] In some embodiments, the sliding window similarity weighting process utilizes a sliding window mechanism to perform a weighted summation of the allocation pattern similarities within multiple consecutive historical time windows to reflect the timeliness characteristics of scene matching. Assuming the sliding window length is L, the allocation pattern similarity sequence between each historical sample scene in the window and the current scene is: The sliding window similarity weight is defined as: The weight coefficient is selected in the form of exponential weighting: The parameter λ controls the degree of weight attenuation. The smaller the value, the higher the weight is given to the adjacent moments, so that the historical distribution pattern closer to the current moment in time has a higher influence.

[0121] In this embodiment, the comprehensive priority fusion process combines the sliding window similarity weighted result with the historical operation efficiency priority information of the historical sample scene to construct a comprehensive scoring function to evaluate the overall timeliness adaptability of the current allocation plan. The historical operation efficiency priority of the historical sample scene is denoted as Q h , represents the historical scene sample h k The corresponding actual operation performance score, such as flight punctuality, transfer efficiency or resource utilization index and other statistical information. Comprehensive scoring function R(M cur ) is defined as: The above scoring function is used to comprehensively score and rank the candidate solutions in the current physically feasible configuration set, and the solution with the highest score is selected to form the final optimal aircraft position distribution matrix. The matrix elements are recorded as: In this embodiment, through the above-mentioned historical similar scene retrieval processing, allocation pattern similarity quantification processing, sliding window similarity weighted processing and comprehensive priority fusion processing, the timeliness pattern matching and evaluation of the physical feasible configuration set are completed, and the generated optimal aircraft position distribution matrix can more effectively match the actual needs of the current operation scenario, and obtain continuous optimization effects in the dimensions of operation stability and service quality.

[0122] Based on the description of the embodiment of the airport stand allocation method based on grid-based spatiotemporal stratification, the present application also discloses an airport stand allocation system based on grid-based spatiotemporal stratification. The airport stand allocation system based on grid-based spatiotemporal stratification can be a computer program (including program code) that runs the aforementioned airport stand allocation method based on grid-based spatiotemporal stratification. Figure 6 As shown in the figure, the airport stand allocation system based on grid-based spatiotemporal stratification can run the following units:

[0123] Acquisition unit 110 is configured to acquire basic airport operation data, including real-time flight dynamics data, aircraft stand physical attribute data, historical operational performance data, historical flight delay records, ground support resource topology data, airport physical topology data, and ground support resource status data. The real-time flight dynamics data includes planned schedules, delay status, aircraft model parameters, and transfer connection requirements. The ground support resource topology data includes structured description data of the physical connection relationships and spatial layout between support facilities. The ground support resource status data includes dynamic monitoring data of the real-time occupancy status, movement trajectory, and service load of support equipment.

[0124] The grid modeling unit 120 is used to perform spatiotemporal correlation analysis and grid modeling processing on the basic data to construct an airport spatiotemporal grid model. The grid modeling processing includes spatiotemporal demand weight calculation and grid optimization fitting.

[0125] An adaptive optimization unit 130 is configured to perform adaptive optimization processing on the airport space-time grid model by coupling multi-level feature analysis with dynamic delay disturbances, thereby obtaining an airport stand distribution matrix. The adaptive optimization processing includes multi-level feature constraint fusion, dynamic adaptive weight adjustment, and low-dimensional reconstruction optimization based on sparse constraints.

[0126] The anti-conflict coupling unit 140 is used to perform dynamic coupling optimization and anti-conflict verification on the airport stand distribution matrix and ground support resource status data to generate an anti-conflict stand configuration set;

[0127] The configuration verification and result output unit 150 is used to perform real-time physical constraint verification and historical performance mode optimization on the conflicting aircraft position configuration set to generate an optimal aircraft position distribution matrix.

[0128] The foregoing description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein and should not be construed as excluding other embodiments. Rather, the present invention can be used in various other combinations, modifications, and environments and can be modified within the scope of the concept described herein through the above teachings or techniques or knowledge in the relevant field. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention are intended to be protected by the appended claims.

Claims

1. A method for allocating airport stands based on grid-based spatiotemporal stratification, characterized in that: The method comprises the following steps: S1. Obtaining basic airport operation data, including real-time flight dynamics data, aircraft stand physical attribute data, historical operation performance data, historical flight delay records, ground support resource topology data, airport physical topology data, and ground support resource status data. The real-time flight dynamics data includes planned schedules, delay status, aircraft model parameters, and transfer connection requirements. The ground support resource topology data includes structured description data of the physical connection relationships and spatial layout between support facilities. The ground support resource status data includes dynamic monitoring data on the real-time occupancy status, movement trajectory, and service load of support equipment. S2, performing spatiotemporal correlation analysis and grid modeling on the basic data to construct an airport spatiotemporal grid model, wherein the grid modeling includes spatiotemporal demand weight calculation and grid optimization fitting processing; S3, performing adaptive optimization processing on the airport space-time grid model by coupling multi-level feature analysis with dynamic delay disturbances to obtain the airport stand distribution matrix. The adaptive optimization processing includes multi-level feature constraint fusion, dynamic adaptive weight adjustment, and low-dimensional reconstruction optimization based on sparse constraints. S4, dynamically couples and optimizes the airport stand distribution matrix and ground support resource status data, and performs anti-conflict verification to generate an anti-conflict stand configuration set; S5, perform real-time verification of physical constraints and optimization of historical performance patterns on the conflicting aircraft configuration set to generate the optimal aircraft distribution matrix.

2. The method for allocating airport stands based on grid-based spatiotemporal stratification according to claim 1, characterized in that: The S2 step includes the following sub-steps: S201, performing spatial correlation analysis on the physical attribute data of the aircraft stand and the topological data of the ground support resources, and generating an aircraft stand interaction network through adaptive scene mapping, wherein the spatial correlation analysis is a spatiotemporal adjacency calculation of the aircraft stand based on multi-scale topological aggregation; S202: Calculate spatiotemporal demand weights based on the stand interaction network and real-time flight dynamic data, and generate a dynamic demand density matrix through asynchronous spatiotemporal grid mapping. The spatiotemporal demand weight calculation is based on weight allocation of flight dynamic priorities and time window conflict. S203, performing spatiotemporal optimization fitting on the dynamic demand density matrix and historical operational efficiency data to generate an airport spatiotemporal grid model. The spatiotemporal optimization fitting is based on the aircraft stand occupancy rate and connection time distribution in the historical operational efficiency data, and the grid density is adjusted using a local sparse enhancement algorithm.

3. The method for allocating airport stands based on grid-based spatiotemporal stratification according to claim 1, characterized in that: The S3 step includes the following sub-steps: S301: Perform multi-level feature analysis and constraint fusion on the airport space-time grid model to generate a hierarchical optimization tensor. This hierarchical feature analysis and constraint fusion includes regional coordination matrix modeling at the international / domestic coordination layer, aircraft type-stand bipartite graph matching constraint generation at the wide-body / narrow-body adaptation layer, and temporal network flow analysis at the transfer connection optimization layer. A multi-dimensional optimization space is constructed through tensor superposition and hierarchical weight fusion. S302: Adaptively adjust the weights of the hierarchical optimization tensor and the flight delay time series data to generate an optimization constraint space; the adaptive weight adjustment is based on time series delay disturbance analysis. The delay time series data is obtained by performing sliding window aggregation and propagation chain inference processing on the current delay status in the real-time flight dynamic data and historical flight delay records; S303, performing adaptive tensor decomposition and low-dimensional solution space reconstruction processing on the optimized constraint space to obtain the airport stand distribution matrix, wherein the adaptive tensor decomposition and low-dimensional reconstruction processing include adaptive sparse constraint application based on dynamic scenes and low-rank representation of dynamic threshold screening.

4. The method for allocating airport stands based on grid-based spatiotemporal stratification according to any one of claims 1 to 3, characterized in that: The S4 step includes the following sub-steps: S401, performing a dynamic coupling feasibility analysis on the airport stand distribution matrix and ground support resource status data, generating a stand-resource joint feasible domain matrix through a spatiotemporal constraint network. The dynamic coupling feasibility analysis is based on conflict detection and capacity matching between the physical topology constraints and the real-time resource load matrix; S402, performing multi-objective dynamic anti-conflict optimization on the joint feasible domain matrix of the stand and resource to generate an anti-conflict stand configuration set, wherein the multi-objective dynamic anti-conflict optimization is based on a linear combination scoring function of the conflict resolution weight, the resource balancing factor and the flight priority vector, and is implemented through sparse solution space projection and local neighborhood search.

5. The method for allocating airport stands based on grid-based spatiotemporal stratification according to claim 4, characterized in that: The S5 step includes the following sub-steps: S501, performing real-time spatial conflict detection on the conflicting aircraft stand configuration set and the airport physical topology data, and generating a physically feasible configuration set through dynamic geometric boundary verification. The real-time spatial conflict detection is based on aircraft stand safety spacing constraints and vehicle path width threshold analysis. S502, performing time-effectiveness pattern matching on the physical feasible configuration set and the historical operating efficiency data, and generating an optimal aircraft position distribution matrix by sliding window similarity weighting. The time-effectiveness pattern matching is based on the optimal allocation pattern extraction and priority fusion under historical scenarios of the same type.

6. The method for allocating airport stands based on grid-based spatiotemporal stratification according to claim 4, characterized in that: The physical topology constraints in step S401 are obtained by constructing an adjacency matrix and performing graph-theoretic connectivity analysis on the airport stand distribution matrix, and the real-time resource load matrix is ​​obtained by performing time series discretization and sliding window aggregation on ground support resource status data.

7. The method for allocating airport stands based on grid-based spatiotemporal stratification according to claim 4, characterized in that: The conflict resolution weight in step S402 is obtained by dynamic statistical calculation of the conflict mark density in the slot-resource joint feasible domain matrix, the resource balancing factor is obtained by normalized analysis of the distribution entropy value of the real-time resource load matrix, and the flight priority vector is generated by static weight assignment of flight attributes.

8. The method for allocating airport stands based on grid-based spatiotemporal stratification according to claim 5, characterized in that: The real-time spatial conflict detection in step S501 is based on the aircraft position safety distance constraint and the vehicle path width threshold analysis, including the aircraft position dynamic geometric boundary construction processing, real-time distance safety verification processing and vehicle path width verification processing.

9. The method for allocating airport stands based on grid-based spatiotemporal stratification according to claim 5, characterized in that: The temporal pattern matching in step S502 is based on the optimal allocation pattern extraction and priority fusion under historical scenes of the same type, including historical scenes of the same type retrieval processing, allocation pattern similarity quantification processing, sliding window similarity weighting processing and comprehensive priority fusion processing.

10. An airport stand allocation system based on grid-based spatiotemporal stratification, characterized in that: The system comprises: an acquisition unit, configured to acquire basic airport operation data, including real-time flight dynamics data, aircraft stand physical attribute data, historical operation performance data, historical flight delay records, ground support resource topology data, airport physical topology data, and ground support resource status data. The real-time flight dynamics data includes planned schedules, delay status, aircraft model parameters, and transfer connection requirements. The ground support resource topology data includes structured description data of the physical connection relationships and spatial layout between support facilities. The ground support resource status data includes dynamic monitoring data on the real-time occupancy status, movement trajectory, and service load of support equipment. A grid modeling unit is used to perform spatiotemporal correlation analysis and grid modeling on basic data to construct an airport spatiotemporal grid model. The grid modeling includes spatiotemporal demand weight calculation and grid optimization fitting. An adaptive optimization unit is used to perform adaptive optimization processing on the airport space-time grid model by coupling multi-level feature analysis with dynamic delay disturbances to obtain the airport stand distribution matrix. The adaptive optimization processing includes multi-level feature constraint fusion, dynamic adjustment of adaptive weights, and low-dimensional reconstruction optimization based on sparse constraints. The anti-conflict coupling unit is used to dynamically couple and optimize the airport stand distribution matrix and ground support resource status data, and perform anti-conflict verification to generate an anti-conflict stand configuration set; The configuration verification and result output unit is used to perform real-time verification of physical constraints and historical performance mode optimization on the conflicting aircraft configuration set to generate the optimal aircraft distribution matrix.

Images